Background, Motivation and Objective Echocardiographic strain (rate) imaging allows quantifying regional myocardial function. In clinical applications, most often, end-systolic strain and peak systolic strain ratehave been used as techno-markers to classify disease. However, as these values characterize the strain (rate) curve at a single time point potentially important diagnosticinformation in the temporal behavior of these curves gets lost. The aims of this study were therefore bi-fold: 1) to build a statistical model enabling a compact representationof the temporal strain (rate) profiles and, 2) to use this compact representation to build a feature vector to detect disease.